RF digital cellular systems utilizing time division multiplexed carriers offer distinct advantages over analog cellular systems in both performance and subscriber capacity. A typical RF digital cellular system divides a given carrier frequency into identifiable time frames which comprise multiple time slots wherein each slot contains the digitized and coded speech (or data) for one traffic channel (one voice). This is referred to as a time division multiplexed (TDM) carrier. Therefore, one carrier frequency may accommodate multiple traffic channels. The Groupe Special Mobile (GSM) Pan-European cellular system, as specified in GSM recommendations available from the European Telecommunications Standards Institute (ETSI) and incorporated herein by reference, is such a digital cellular system. As typical with conventional cellular systems, the GSM system relies on geographic reuse of frequencies to achieve the large subscriber capacities. The use of reuse clusters involving various antenna configurations and patterns is well known. In the GSM terminology, the concept of a cell involves radio service to particular geographic area and is not related to the location of the equipment. Thus, various sectors of an antenna comprise different GSM cells.
The GSM system requires frame synchronization for all carriers used in the same cell, i.e. a frame of one carrier frequency must be synchronized with a frame of all of the other carrier frequencies transmitted from the same cell. This ensures the ability of establishing the frame number for all the carriers of the cell based on information from a single carrier which provides this information, called the BCCH carrier. Each cell site uses a plurality of substantially identical multiframe signalling structures on multiple carriers and each carrier is assigned, at most, 8 voices or traffic channels. FIG. 1 illustrates typical GSM timing for one carrier. Each frame (100) contains 8 slots (S1-S8) with each slot representing one voice channel. In addition, GSM requires that a given TDM frame (100) must have slots which have the same representative coded information from each input voice, that is the constituents (algorithmic entities) of information for all users (slots) within a frame is essentially identical. The requirement of substantial identify of information within slots of a frame provides that each voice channel represent a frame number dependent communication resource for each user of the system.
For example, a given frame (100) must contain the same real time portion of speech coded information from each voice for each slot. Here, a 20 msec block of raw speech is taken from each of the traffic channels (105) and sampled at 8000 Hz A-law PCM and then transcoded (110) (essentially low bit rate encoded) by digital signal processors (DSP's) using RPE-LPC/LTP (the GSM specified speech coder technique), or other suitable coding technique. The transcoded speech for each 20 msec block then undergoes redundant coding (115) via convolutional coding techniques and is then partitioned and interleaved in slots (100) forming the coded speech bursts. This process is done for each of the 8 traffic channels. Since there is a stringent delay specification between the time a raw speech block is sampled to when transcoding by the DSP is to begin, one digital signal processor is required for each traffic channel.
A TDM system that requires same frame synchronization (Frame 1 of any cell is synchronized with the Frame 1 of all cells) between all surrounding cells, to produce a method for rapid handoffs between the subscriber and an adjacent cell. The knowledge of this synchronization by the subscriber allow him to determine his absolute distance from the target cell and adjust his transmissions accordingly to minimize any handover interruption.
However, using asynchronous, or particularly the frame synchronous TDM system, as generally described above, creates numerous problems affecting overall system performance, cost, layout, and reliability, particularly in the base sites or switch centers where the speech transcoding takes place and the framing is controlled. Requiring the same algorithmic outputs from exactly overlapping real time portions of each voice to occupy the same frame, combined with requiring restrictive timing delays for transcoders necessitates providing one DSP for each traffic channel (voice).
This compels the need for 8 DSP's for each carrier used. Requiring such substantial numbers of DSP's directly impacts cell and base site layout (size of cages to house these devices and supporting circuitry) complexity, cost, and performance.
Another problem concerns worst case handoff latency in the same frame synchronization system. The existing GSM system uses a broadcast control channel (BCCH) to communicate control and access information to all subscribers units within its geographical receiving area. Handoff latency is measured as the time it takes for a subscriber to synchronize to and decode adjacent cell (potential destination cells for the handoff) BCCH carrier identifications. The BCCH bursts needed to accomplish this occur relatively infrequently in time. Delay in this process has the potential of causing audio degradation and dropped calls. Although the subscriber unit in such a system typically assists in the handoffs by monitoring the signal strength or distance of surrounding cells and maintains a table of the best candidates, typically, the subscriber unit can only decode this information during an idle frame which occurs every multiframe. With all surrounding cells synchronized, the BCCH identification for all adjacent cells is always transmitted at the same time. Therefore the subscriber only receives the BCCH identification from any adjacent cell after a sizeable delay due to the need to have the signalling multiframe slide significantly relative to the traffic multiframe.
Accordingly, there exists a need for an RF cellular TDMA communication system that provides efficient speech transcoding for multiple traffic channels and also improves handoff latency.